Headphone calibration software (7 April , 2021)



What is headphone calibration?

Over the past decade, impressive advances in digital signal processing (DSP) technology have resulted in new techniques for controlling the spectral coloration of studio monitors and headphones. Taking into account the cost and complexity of the acoustic treatment of studios and, in today's case, non-purpose-built mission-critical listening rooms (studios in bedrooms, etc.), software developers such as IK Multimedia and Sonarworks have created cool software products to hypothetically " neutralize".


ARC vs Sonarworks vs Nura IK Multimedia's

ARC software and Sonarworks' Reference 4 (among others) allow the user to measure with a calibration microphone, which then generates an impulse response that represents the spectral attributes of the room. With this information, the software can feed back directly to the output of the monitoring system in order to deconstructively superimpose the non-planar frequency response. Manufacturers like Nura from Melbourne have developed headphones that automatically measure otoacoustic emission signals (OAE) in order to represent the transfer function of the ear canals of the listener (and to process the digital signal from the headphones).

Sonarworks Reference 4 offers an additional feature over competing products: an integrated set of calibration profiles for many popular brands of high quality and consumer reference headphones. Given the widespread use of music critical listening rooms outside the destination, this seems like an attractive alternative to expensive room acoustics. Compared to the Reference 4 room measurement calibration system or Nura's OAE system, these headphone calibration profiles do not require user measurements and therefore do not require an expensive measurement microphone or complex built-in surveillance system.


Destroying myths about the "ideal flat curve" and in difficult words - the ideal frequency response

The built-in headphone calibration presets sound like a great tool for the home composer, but as with most digital signal processing, there is almost always additional, potentially undesirable, effects on the digital waveform. In addition, it is a generally accepted fact that even the position of a pair of headphones on the head can introduce variability in frequency response, not to mention fluctuations caused by differences in the size of the ears of the headphones and the ears of each earphone Individually. We ran a series of tests to determine the variability in frequency response due to these factors.

The headphones were placed on a Neumann KU100 dummy head (a kind of crash test dummy with microphones in the ears). The frequency response was measured using a sinusoidal sweep in an anechoic chamber. With each measurement, the headphones were removed and then put back on the wrong ears. The following graphic shows the change in a pair of Hifiman He400i with ten resets per ear (blue graphic - left ear, red - right).



Not only is there a large discrepancy between the measurements, but there is an even larger discrepancy between the left and right drivers. Against this background, how is it possible that a headphone calibration profile that is suitable and recommended for everyone produces the ideal frequency response in our particular case? We can conclude that the profiles shown are somewhat averaged and each pair needs to be measured individually (by the way, you can buy pre-calibrated headphones from Sonarworks and look for models and prices on the website).

For comparison, the reference calibration curve provided by Sonarworks for the same headphones looks like this ...



Not exactly the same?


Raise the bass or not raise it?

If you look at the documented frequency response of closed or open headphones, it becomes clear: Small headphone drivers, dynamic or planar magnetic, really cannot reproduce any sound in the lower frequency spectrum. As most people in music production know, it's much easier to get low-frequency sound with a larger diaphragm and deeper speakers than it is with shallow little drivers.



For one of my favorite headphones, the Sennheiser HD600, Sonarworks corrects 20-50 Hz to 10 dB. Anyone who has ever tried creating too much low frequency sound through a portable bluetooth speaker knows how to do it. Why should this be any different for headphone drivers? In fact, non-linear behavior or distortion occurs due to an overload of the 20-50 Hz range during the sound reproduction by the driver.



In the picture above you can very clearly see the main peak (far right) and then the first, second, third etc. harmonics seen from right to left. This is a direct result of the low frequency signal being overloaded during calibration.


What does that mean?

As producers, we're used to using distortion and saturation as creative tools. If we were to analyze the frequency spectrum before and after using a distortion module or external device in the signal, you would very quickly notice the appearance of harmonic peaks after applying the distortion. Now think about the overall signal coming from your headphones.

The goal of calibration technology is to correct peaks and valleys to achieve a smoother frequency response. However, this is a direct result of the amplification of low frequency valleys where we inadvertently create new medium and high frequency content. Nonlinear content is difficult to predict and difficult to repair, affecting your carefully crafted audio.

Then what are the real solutions? Software lovers and purity advocates will have different opinions about the best alternatives. Obviously, processing the headphone drivers in this way creates unwanted signal artifacts. While the lower and upper ranges of the human auditory spectrum have variation and possible distortion when playing audio with headphones, you can only aim for the mids to reduce the effects of natural coloring, not the highs and lows that are problematic .

This appears to be the only frequency range that is reasonably consistent and unaffected by other products that can be improved through the use of digital signal processing technology. For this purpose, the reference 4 offers the possibility of adapting the reference curve according to your individual preferences. Create your own curve by adjusting the Bass Boost and Slope settings until the minimum processing is applied to ≤ 100 Hz and ≥ 6000 Hz. However, with this option it is very difficult to get the correction curve exactly how you want it, which brings us to the end point.


Then why do you even need it?

Why process audio in headphones at all to achieve a flatter frequency response? As budding producers, we should study the speakers, explore the listening room and know the advantages and disadvantages of the acoustic environment we are immersed in. So why should we create a whole new, colorful world to retune our ears? Room modes (studio monitor calibration) can produce large peaks at certain frequencies that mask other frequencies, cause problems in our mix, etc.

The main difference, however, is that room modes can be measured with calibration microphones, which measurements are fairly accurate at the customer level. In the reference headphones area, however, the peaks and valleys we are dealing with are not measurable. We rely on average values that do not take into account differences between headphone design, ear shape and head position. We also know the Hifiman HE-400i has an odd 2km jump, and that doesn't make us love it any less.

With that said, we're going to be making an outdated opinion - you need to stop messing with headphone calibration profiles and just listen to as much music as you can in your headphones until you become one with them. Compare your work with other artists of the same style, take an A-B test in selected areas, and remember that the end user of your work is almost certainly not listening to music with reference headphones worth $ 2,000.

The idea of a perfectly flat turn is fun, but it probably creates more problems than it's worth. What do you think of the headphone calibration? Share the article on social media and tag our profiles with thoughts on the subject.

Headphones VST (Intended for those who just want to / have to mix more through headphones.)

Ambi Head by Noise Makers - Ambisonics Binaural VST
Beyerdynamic - Virtual Studio – Inspired by Headzone VST
Sennheiser - Ambeo Orbit VST
ToneBoosters - Morphit VST
Sennheiser - dearVR MICRO VST

 

Last edited: May 5, 2021

I doubt that this is possible. To calibrate a HP you need to put them on a dummy head (with special in-ear mics).

I haven't heard of a free HP calibration software either.
Additionally this software would need a preset (EQ correction curve) for your HP and I doubt that anyone has one for this HP.

My suggestion would be to use either a system wide EQ or an EQ in your DAW's mainout (don't forget to switch it off for rendering) and raise the low freqs. Make some test adjustment (different settings) and compare them to other systems to find out which adjustment is most compatible.

 

You have closed back isolation monitors, they are always going to be de-amplify bass mid lows and bass all that energy trapped in the cups is just going to push your inner ear with more force because you get pressure resonance waves de-emphasis.

It's the same concept used in some car intake systems to get over atmospheric pressure in a cylinder in a naturally aspirated engine. The same thing happens in your closed headphone cup but in reverse with high energy low frequency waves + HRTF. Open back headphones or semi-open will give you more natural bass but only to a certain extent. At least you are not going to get cancelation from pressure wave collisions in a mostly closed system.

As far as correction graphs there are some headphone specific forums that will do a frequency response on a test head and HRTF calcs on request if you get lucky and someone feels generous. Otherwise you have to invest in a known quantity.

 

Dirac Live Studio is good for calibrating studio monitors.

https://www.audiosciencereview.com/...-sonarworks-reference-4-studio-edition.18607/

 

Last edited: Jul 25, 2021

@BEAT16 You've touched an interesting subject that I just had to deal with recently.
Basically, many of the correction plugins are either parametric equalizers fed with the appropriate settings or realtime convolution engines that are fed with a "reverse impulse response" aiming at correcting the deviations from a linear response in a certain headphones model.
Data of thousands of headphones models that have been measured are freely available and a simple start is to use a parametric EQ with enough bands to apply them.
And it's true, radical settings in the bass range would theoretically give us a better response but these settings can be extreme enough to limit monitoring to unacceptably low levels.
What I had to do (with the EQ curves at least) is to apply all fixes except the ones in the lower bass range which I had to tame down manually in order to make them work at my typical listening levels, and that's especially true with headphones that don't have strong bass response.

 

What headphones were you using? Many pro's use planar headphones for their deep fast bass, with Audeze LCD-X being a popular choice.

 

Last edited: Jul 25, 2021

A DT770/80 Ohms, mainly because of the comfort and closed back build.
Originally unusable for monitoring, after the EQ treatment it's quite an acceptable set.

The LCD-X are fantastic but they give you a level of reproduction that hardly any consumer uses.

 

I'm using ATH-M50 & HD600 with Morphit and CanOpener. I am tempted by the LCD-X as I don't have a great acoustic space to work in. I use E5 XT monitors to check stuff, but I mostly work with headphones for tracking and mixing. However, you are correct in saying that this an interesting subject, as we all attempt to defy the laws of physics

BTW Dirac Live Studio have a 14 day demo and it's really good, I'd say its better than Sonarworks & IKM Arc for speaker correction.

 

Last edited: Jul 25, 2021

@Zenarcist Interesting subject indeed, I was surprised how effective a simple 10 band PEQ was. Now the 770 sounds like it should have sounded in the first place.

Oh and btw, here's where I got the settings from:
https://github.com/jaakkopasanen/AutoEq/blob/master/results/INDEX.md

I would say that most headphones on the market are included.

 

I can recommend a book from Stavrou: Mixing with your Mind. Inside, he has described quite well that the ears are a great measurement tool too! There are certain methodologies on how to set up speakers without measurement software or microphones. The most important and mostly underestimated point is, to set up the distance between the speakers correctly. Sometimes or even mostly its not just the perfect triangle! To set up the speaker distance you need to let the speakers face each other, listen to a good mono track and move your head in between the speakers to hear the phase. You will reach a point where you can hear the flesh of the singer for example, a matter of centimeter or less sometimes. Imagine a rubber band between them. Sometimes its too loose and sometimes to tight.

The other point is, that you need to know your speakers very well, then you dont need any calibrations, because you know what you do. Everybody who starts has problems with bass and so on, calibration software cant help with that as much as correct speaker placement and room treatment can do. More important it is that your mixes are rocking your speakers, like a guitarist learns how to rock the amp. Takes years to learn.


@OP: Headphone calibration is awful in my experience and wont help with the mixes. I know that many people think its super important. The problem with all of that is, that your brain will compensate it anyway, so I would say its not important how you calibrate them. You should own a pair of good headphones that you know well, you listened to the high quality records you know well, have listened to them on many other systems and after a few hundered hours you know what to trust.

The combination of carefully placed speakers and a good pair of headphones, both you know well, is the key. No measurement software can help you with that as much as your brain is able to help you with the human build-in compensation. The build-in one is also for free!

 

My ATH-M50 improved a lot with EQ.

Here is some more general info ...

In depth headphone reviews with measurements and suggested EQ correction curves:

https://www.audiosciencereview.com/forum/index.php?pages/HeadphoneReview/

Oratory1990 EQ correction curves:

https://www.reddit.com/r/oratory1990/wiki/index/faq

https://www.reddit.com/r/oratory1990/wiki/index/list_of_presets

 

Last edited: Jul 25, 2021

Since I started using Morphit & CanOpener I cant use my headphones without them. I listen to a lot of reference tracks this way and its a lot more enjoyable. I agree with you about using near field speakers for cross reference.

 

Last edited: Jul 28, 2021

Well, I have a pretty dead (we are talking Anechoic almost as the RT60 is 150ms at 50Hz and from 100 to 8k, it is 100 to 90ms and the room is extensively treated) room.
Here is what I found using double blind tests with both audio professionals and normal human beings (5-6 people). Double blind tests are usually pretty reasonable.

So here's what I did. I chose some songs that I know. I monosummed it in the DAW. Then sent that signal to Sonarworks.
Now, everytime, with Sonarworks on, the stereo image got widened.
What does this mean?
Sonarworks is creating fake stereo width.
Well how?
Sonarworks applies more than needed L-R offsets (EQ spectrum and level).
No matter how many times I calibrate using Sonarworks, this is the case everytime.

Explanation?
Well, no matter how precisely you measure, you will always be 1-2 degrees off or even 1/16th inches off. So unless during calibration, an AI Robot holds the mic, this will always be the case.
Now Sonarworks recommends, handheld Mic. This means mic will always capture humanbody reflections.
I have done the calibration process with handheld mic and with mic on a stand but everytime double blind test did yield the same.
With Sonarworks enabled, the mono signal is getting slightly more width when played back on the speakers compared to the same mono signal being played by the speakers with SW off.

So now I do not trust SW monitor calibration. I don't know about the headphones though.

P.S before SW treatment, I have tested my room using REW.

 

My DT-880 (completely unusable without correction), M50X and even my HD650 improve greatly with Sonarworks. I guess it's not about achieving a flat response, but more about removing the biggest lies our headphones tell us.

 

Last edited: Jul 26, 2021

i still use mathaudio auto eq to this day (since 2012) with my shp-9500/s 6hz-35000 khz.. kicks sum ane-uss. might be somewhat difficult to find though. the modern edition ''headphone eq'' is far inferior. thats why i still use autoeq R2R.

 

I need to use something like this, years ago I started to use a phone app, don't remember the name but was something like "normalizer" didn't work well with my new phone so I stopped used it, but it was very effective in the old one.

It generated pulses in different frequencies so you did the calibration manually based on your own hearing/headphone acoustics, then the app run in the background equalizing any audio player in the phone, the change was amazing.

Anyway now I am thinking that this is not right for producing/mastering or not all the time as any problem with the hearing for example the issues of higher frequencies after getting too old, I guess is a gradual thing and also not with it's own right/left or frequencies abnormalities, but I guess that also can be compensated with the right calibration.

 

thank you very interesting

 

Low THD Headphones under 1% below 100Hz:
Audeze LCD-24 (Open-back)
Neumann NDH 20 (Closed-back) [Replace suede earpads with velour
or hybrid earpads to reduce high mids scoop]

Low THD Headphones under 2% below 100Hz:
Sennheiser HD 280 Pro MKII (Closed-back) [Replace fake leather earpads with velour
earpads to reduce high mids scoop]

Low THD Headphones under 4% below 100Hz:
Audio-Technica ATH-R70x (Open-back) [Replace velour earpads with hybrid
earpads to add lows and reduce high peaks]
Shure SRH1540 (Open-back)
Sennheiser HD 300 Pro (Closed-back) [Replace suede earpads with velour
earpads to reduce high mids scoop]

Hybrid earpads for more sub
Velour earpads for more treble
Leather earpads for no leakage

Also low thd headphone are best for eq modification also an earpad change sometime is essential if sound sound is your aim.

 

Haven't been around lately, so if this has been covered, no flame, please.
Just wanted to throw another dog into the fight.
I just demo'ed Sienna Free, and found it very interesting. Not "exactly" a headphone calibration tool, but rather a virtual mixing room in headphones type thing.
The interesting part in regards to the thread is Acustica has around 180 different headphone types they've "corrected" for so in the headphones you are using, it should sound like you are in that room. Magically, they have my Samson SR-850's that are very non-expensive open back headphones as well as my AT-M50X. They also have Beats, Airpods, etc. for listening on the go.
Hell, this is free until tomorrow, I think, so it's worth a checkout to anyone who is interested in this kind of thing.
My experience, after trying it out with the Airpod setting using my Cambridge Audio Melomania wireless earbuds, is even though the headphone model was wrong for what I was using to listen with, I could hear subtle changes to compression and eq'ing better. It is supposed to be a tool to help with system translation, primarily.
Much of the feedback I've read indicates that you really have to do some reference checking and listen to alot of material in order to get used to what it's doing. I suppose to train your ears to what a flat response is supposed to sound like. Don't just jump in and start trying to mix with this.

tldr- check out Acustica Sienna, thank me later.

 

Last edited: Jul 30, 2021